- © 1999 by American Society of Clinical Oncology
Phase I Study of Liposomal Vincristine
- K. A. Gelmon,
- A. Tolcher,
- A. R. Diab,
- M. B. Bally,
- L. Embree,
- N. Hudon,
- C. Dedhar,
- D. Ayers,
- A. Eisen,
- B. Melosky,
- C. Burge,
- P. Logan and
- L. D. Mayer
- From the British Columbia Cancer Agency, Vancouver, University of British Columbia, and Inex Pharmaceuticals, Burnaby, British Columbia, Canada.
- Address reprint requests to Karen Gelmon, MD, British Columbia Cancer Agency, Vancouver Centre, 600 West 10th Ave, Vancouver, British Columbia V5Z 4E6, Canada; email kgelmon{at}bccancer.bc.ca
Abstract
PURPOSE: A phase I study of vincristine encapsulated inside 120-nm-diameter distearoylphosphatidylcholine-cholesterol liposomes was performed. The primary objectives were to determine the maximum-tolerated dose (MTD), recommended phase II dose, toxicity, and pharmacokinetics of liposomal vincristine (ONCO-TCS).
PATIENTS AND METHODS: Twenty-five patients with histologically confirmed malignancies were enrolled and assessable. Vincristine doses were increased from 0.5 mg/m2 to 1.0, 1.5, 2.0, 2.4, and 2.8 mg/m2 with cohorts of three or more patients per dose level. A total of 64 courses of ONCO-TCS were administered intravenously once every 3 weeks. The pharmacokinetics of total vincristine content in plasma were determined using a high-performance liquid chromatography method.
RESULTS: Patients were treated with vincristine doses up to 2.8 mg/m2; however, 2.4 mg/m2 was defined as the MTD and 2.0 mg/m2 as the phase II recommended dose. Pain and obstipation were the dose-limiting toxicites. Other toxicities were fever, rigors, fatigue, myalgias, and peripheral neuropathy. Hematologic toxicity was mild. All patients who were treated with doses above 1.5 mg/m2 received in excess of 2.0 mg of vincristine, with doses as high as 6.2 mg. One partial response was seen in a patient with pancreatic cancer. Tumor response not meeting partial response criteria was seen in two other patients. Pharmacokinetic studies revealed significantly elevated concentrations of total vincristine, but parameters varied and were not directly correlated with toxicity or response.
CONCLUSION: The ability to administer elevated doses of vincristine, as well as indications of efficacy, suggests that ONCO-TCS warrants further clinical investigation in a phase II setting.
LIPOSOMAL ENCAPSULATION of anticancer drugs has been studied extensively both in the laboratory and in the clinic, with reports of increased plasma drug concentrations, improved tumor delivery, decreased toxicity, and increased efficacy for a variety of cytotoxic agents.1,2 Liposomal formulations of the anthracyclines daunorubicin and doxorubicin have been approved for the treatment of Kaposi's sarcoma, and other liposomal anticancer therapies are in various stages of clinical development.3-5 However, the benefits associated with liposome applications in the area of anticancer drug delivery have been related primarily to reduced toxicity associated with the encapsulated form of the drug. Significant improvements in antitumor potency have been far less frequent, perhaps because of the fact that most efforts to date have focused on anthracyclines, which exhibit only a modest dependence of cytotoxic potency on the duration of tumor drug exposure.6
The properties of increased circulation lifetimes and increased tumor drug accumulation in conjunction with decreased toxicity seem particularly well suited for cell cycle–specific drugs such as vincristine. Vincristine is an extremely potent agent whose cell cycle specificity arises from its tubulin depolymerization action, which arrests cell growth in metaphase.7,8 Various cell culture studies have shown that the antitumor potency of vincristine increases by several orders of magnitude when the drug exposure to tumor cells is prolonged.6,9 Given this information, it may be expected that the duration of drug exposure in plasma and tumor sites is critical in determining the degree of therapeutic effect. However, attempts to overcome the rapid elimination of vincristine from the blood after intravenous administration through continuous infusions or increase the dose above the traditional 2 mg/patient cap have resulted in a significant increase in unacceptable neurologic toxicities.10-12 In this context, we have undertaken a series of investigations to develop and optimize the formulation and biologic properties of liposomal vincristine with the objective of greatly enhancing the drug's antitumor activity without exacerbating toxic side effects.
Preclinical evaluations demonstrated that altering the physical properties of liposomal vincristine formulations results in profound changes in efficacy while affecting only modest changes in drug toxicity characteristics. Increasing the retention of vincristine inside liposomes after intravenous administration led to dramatic increases in antitumor potency, particularly when compared with the efficacy obtained with free vincristine.13-15 This is consistent with the steep dependence of vincristine antitumor potency on the duration of drug exposure,9,13 as well as the fact that retention of vincristine in most tissues, including tumors, is rather poor.9 It seems that the ability to prolong the exposure of vincristine in vivo is more important than peak drug concentrations, and this effect is maximized when small (100 nm) liposomes are used. This relates to the ability of such liposomes to selectively extravasate in the leaky vasculature associated with tumor growth.16-18 Formal toxicology studies of a 120-nm-diameter distearoylphosphatidylcholine/cholesterol (DSPC/Chol) liposomal formulation of vincristine in rodent and canine models revealed a similar profile of target organ toxicity for both free and liposomal vincristine, with an increased maximum-tolerated dose (MTD) for the liposomal formulation being observed in mice.19
In view of the dramatic improvements in antitumor potency for vincristine accompanied by a potential decrease in toxicity, a phase I open label, dose-escalating clinical trial of a DSPC/Chol liposome formulation of vincristine sulfate (ONCO-TCS) was initiated at the British Columbia Cancer Agency, Vancouver Center. The objectives were (1) to assess the safety and toxicity profile of this formulation, (2) to determine the MTD as well as the phase II recommended dose given by an intravenous infusion once every 3 weeks, and (3) to study the human pharmacokinetics of vincristine when administered in a liposomal vehicle (Onco-TCS) during and after a single 1-hour intravenous infusion. In this article, we describe the toxicologic and pharmacokinetic behavior of ONCO-TCS administered to patients and discuss the results in the context of previous reports describing the behavior after free vincristine administration.
PATIENTS AND METHODS
Eligibility Criteria
Patients were eligible if they had a histologically confirmed malignancy that was refractory to conventional forms of cancer therapy or a malignancy for which there was no effective standard chemotherapy. Inclusion criteria also included the following: age greater or equal to 18 years; a life expectancy of at least 12 weeks; an Eastern Cooperative Oncology Group performance status of 0 to 2; adequate bone marrow function, as defined by an absolute granulocyte count of more than 1,500/dL, a platelet count of more than 100,000/dL, and a hemoglobin count of more than 8.0 g/L; adequate hepatic function, as defined by a total bilirubin concentration of less than 1.25 times normal; and adequate renal function, as defined by a serum creatinine concentration of less than 1.25 times normal. Patients were not permitted to have evidence of preexisting neurologic dysfunction. Patients were required to be available for frequent visits and treatment at the Vancouver Clinic of the British Columbia Cancer Agency. All patients gave written informed consent to the study, including the pharmacokinetic sampling.
Preparation of ONCO-TCS for Administration to Patients
ONCO-TCS (0.16 mg/mL) is a three-part formulation consisting of DSPC/Chol liposomes for injection (100 mg/mL), sodium phosphate for injection (14.2 mg/mL), and Oncovin injection (vincristine sulfate for injection; Eli Lilly Canada, Inc, Scarborough, Ontario, Canada). The DSPC/Chol liposomes and sodium phosphate used in this study were prepared by the Investigational Drug Program at the British Columbia Cancer Agency (Vancouver, BC, Canada). Oncovin was obtained from Eli Lilly and used without further modification. The encapsulation procedure involved adding 1 mL of vincristine sulfate for injection (1 mg/mL) and 0.2 mL of DSPC/Chol liposomes for injection (100 mg/mL) to a sterile vial and mixing. Sodium phosphate for injection (14.2 mg/mL) buffer solution (5 mL) was then added. The mixture was heated for 10 minutes at 63°C (60°C to 65oC) in a water bath and mixed. Pharmacists performed the encapsulation procedure just before ONCO-TCS administration. The formulation was maintained under strict quality control, which included testing both the liposomes and the final ONCO-TCS. The parameters have previously been described.20The complete quantitative formula of ONCO-TCS is provided in Table 1.
Quantitative Formula of Vincristine Sulfate Liposome Injection (0.16 mg/mL)
Study Treatment
Vincristine sulfate liposome injection was administered on day 1 of each 21-day cycle as a 1-hour infusion. All patients were admitted for the first dose of the drug to facilitate pharmacokinetic sampling. Subsequent doses were in the outpatient setting. Chemotherapy was repeated every 21 days for at least two cycles. Although no standard premedication was given, nausea and vomiting were treated with antiemetics as needed and fever was treated with antipyretics.
Study Design
Cohorts of at least three assessable patients were assigned to one of six escalating dose levels of vincristine sulfate liposome injection (ONCO-TCS 0.5, 1.0, 1.5, 2.0, 2.4, and 2.8 mg/m2). An individual study patient was treated at the same assigned dose level of ONCO-TCS throughout the study. For safety, the next dose level was not opened until at least three patients received two doses of ONCO-TCS and were assessed for cumulative neurotoxicity. A dose was considered to be assessable 3 weeks after drug administration. Dose-limiting toxicity was defined as greater than grade 2 nonhematologic toxicity, using the National Cancer Institute of Canada Clinical Trials Group (NCIC-CTG) common toxicity criteria. Patients with stable disease and nonhematologic toxicity less than grade 2 were continued on chemotherapy. Patients were withdrawn from the study f toxicity greater than grade 2 occurred, if there was evidence of progressive disease, or if the patient requested.
Definition of MTD, Toxicity Evaluation
When at least one patient experienced greater than grade 2 nonhematologic toxicity at a given dose, a total of six patients were entered at that dosage level. If two patients experienced grade 3 or 4 toxicity, accrual ceased at that level and a total of six patients were studied at the next lowest dose level. When there was an unacceptable number of patients with grade 3 toxicity, the principal investigators discontinued dose escalation. Neurologic toxicity criteria were expanded to include significant abnormalities on electromyelography (EMG), which required a total of six patients entered at that dose level. If another patient developed evidence of significant neurologic dysfunction on EMG testing, dose escalation ceased. A single neurologist interpreted the EMG findings, and significance was determined according to the NCIC-CTG toxicity grading.
If greater than grade 3 or 4 granulocytopenia occurred in two or more patients on any dose level, three more patients were entered at that dose step. When one of three of these subjects experienced grade 3 toxicity, dose escalation ceased. Once the MTD had been determined, the recommended dose was defined as one dose level lower than the MTD.
On-Study Investigations
Patients received a complete physical examination at enrollment and every week while on the study. Also, at each visit their Eastern Cooperative Oncology Group performance status and weight were recorded. Blood work, including hematology (hemoglobin, WBC, differential, and platelet counts) and chemistry analyses (of electrolytes, creatinine, calcium, glucose, albumin, bilirubin, alkaline phosphatase, gamma-glutamyltransferase, AST, and lactate dehydrogenase levels), was also performed weekly. All patients had a full neurologic assessment, including an examination by the participating neurologist and EMG within 2 weeks of starting the study, after three treatments, and at the end of their participation. Additional neurologic studies were repeated at any time if clinically indicated. Tumors were assessed radiologically before patients were enrolled and every 3 weeks if initial tests were positive. An ECG was done before enrollment and before every 3-week chemotherapy infusion.
Response Evaluation
Tumor response definitions were based on World Health Organization criteria. A complete response was defined as the complete disappearance of all disease. A partial response was defined as a reduction of 50% or more in the product of the mass diameters at completion. Both responses required a duration of more than 4 weeks with a confirmatory scan or examination. Progressive disease was defined as a 25% increase in the product of the diameters of disease. Stable disease was defined as a less than 50% reduction in disease without evidence of a 25% increase. Patients without bidimensional disease were not assessable for response. All other patients were regarded as having stable disease. The progression-free interval was the period from the time of best response to evidence of increasing disease or new lesions.
Pharmacokinetic Sampling
Pharmacokinetic testing was done on the first cycle of ONCO-TCS. Blood from a superficial vein in the arm contralateral to the site of infusion was collected aseptically for vincristine determination in a 7-mL Vacutainer purple top (EDTA) tube. Blood samples were collected before treatment (two samples), during the infusion at 20, 40 and 60 minutes, and after infusion at 15, 30 and 60 minutes and 2, 4, 8, 12, 24, 48 and 72 hours. Plasma vincristine concentrations were determined by high-performance liquid chromatography (HPLC), using a validated analytical method that has been described in detail.21 A standard curve of peak area ratio values (vincristine to internal standard) versus vincrisitine concentration from the calibration standards was used for quantitation. Initial pharmacokinetic parameter estimates were calculated, and an iterative least squares estimation was carried out with PCNONLIN V4.2 (SCI Software, Lexington, KY) for one-, two-, and three-compartment open models with constant intravenous input and first-order output (PCNONLIN models 2, 10, and 19). Data from each patient were evaluated with these three models. The model that most accurately described the data was determined by the Akaike criteria provided by PCNONLIN as well as an overlay of the PCNONLIN model and the individual observed concentration values versus time for each patient. Analysis of variance and one-tailed t tests were used for statistical evaluation of the pharmacokinetic data. Parametric Cmax values were determined from plasma concentration versus time curves for individual patients. Area under the curve (AUC) values were determined using the trapezoidal rule from initiation of treatment to the last time point that vincristine could be quantified.
RESULTS
Patient Characteristics
Twenty-five eligible and consenting patients with advanced carcinoma were recruited into the study. Their clinical characteristics are listed in Table 2 and reflected a distribution of tumor types typical for open-label phase I clinical studies such as this. All patients were assessable for safety. Three patients received only one dose of the ONCO-TCS and were considered unassessable for the response analysis, leaving 22 patients assessable for response.
Summary of Patient Characteristics
Drug Delivery
Six steps in dose escalation of ONCO-TCS (0.5, 1.0, 1.5, 2.0, 2.4, and 2.8 mg/m2) were completed over the course of the clinical trial. The dose was calculated as milligram of vincristine sulfate per meters squared of body-surface area. Twenty-five patients received a total of 64 ONCO-TCS infusions. In the majority of patients, dosing remained constant throughout the study. However, in two patients, the dose was modified because of toxicity. One patient received the first two infusions at the 2.4-mg/m2 dose level, a third infusion at the 2.0-mg/m2 dose, and a fourth at the 2.2-mg/m2 dose because of toxicity at the higher dose. Similarly, another patient received his first infusion at the 2.8-mg/m2 dose level and two subsequent infusions at the 2.4-mg/m2 dose level. Two patients, one at dose level 1.0 mg/m2 and another at 2.8 mg/m2, received one infusion only and died while on study, one of progressive disease and one of possible toxicity. Table 3 is a summary of patient dosing by level, number of ONCO-TCS infusions, actual dose delivered, and dose reductions. Table 3 includes patients whose dose was changed from their initial dose and demonstrates that for all patients administered doses of 1.5 mg/m2 or higher, the total amount of vincristine delivered was significantly in excess of the 2-mg vincristine cap currently used in clinical practice. This value increased to levels as high as 6.2 mg for the 2.8-mg/m2 dose level, which is 3.1-fold higher than for conventional vincristine formulations.
Planned Doses, Dose Escalation, and Doses Delivered (N = 25)
Toxicity
The toxicity of ONCO-TCS can be divided into symptoms during the infusion and toxicities occurring later, including biochemical and hematologic effects. The dose-limiting toxicities were grade 3 myalgias, constipation, and peripheral neuropathy at the 2.8-mg/m2 dose level. As well, at that level there was one death attributed to grade 4 neutropenia and thrombocytopenia, described below.
Early hypersensitivity symptoms were frequent but generally well tolerated. Fever was reported in 60% of the patients during the first infusion and was recorded as grade 2 in all cases, except for one grade 3 fever at the time of the patient's second infusion at the 2.4-mg/m2 dose level. The fevers were transient, and no anticipatory medication was given before the first dose of ONCO-TCS. Four (27%) of 15 patients with a fever on the first cycle had one febrile episode at one subsequent infusion. Rigors, headache, and flushing were each seen on three occasions during the infusion, whereas diaphoresis, tachycardia, nausea, and anxiety were seen less frequently. Pain, described as generalized myalgias and arthralgias and not limited to the intravenous site during the infusion, was seen on two occasions, one time each at the 1.5 and 2.0 mg/m2 level.
Patients were monitored weekly for toxicity, which was graded according to the NCIC-CTG common toxicity criteria. The results are summarized in Tables 4, 5, and 6. As expected, the toxicity at the lower doses was minimal and no grade 3 toxicity was reported until the 2.0-mg/m2 level. At the 2.4- and 2.8-mg/m2 dose levels, 50% of the patients had grade 3 or greater toxicity. The dose-limiting toxicity of myalgia was described as severe muscle pain occurring a number of days after the infusion and requiring narcotic analgesia. Constipation was a problem that was mild at the lower dose levels (1.0 to 2.0 mg/m2) but became severe at the highest levels. Likewise, peripheral neuropathy was mild at the lower dose levels but was seen in over 50% of the patients on the higher doses with increasing severity. The patient on the 2.8-mg/m2 dose who received 6.2 mg of ONCO-TCS experienced grade 3 myalgia.
Toxicity by Worst Grade on Study (N = 25)
Toxicity by Worst Grade for Patients on Doses of 2.0 mg/m2 or Greater
Most Common Nonhematologic Toxicities
Mild anemia was seen in the trial and was attributed to both the advanced disease state of patients and the drug. Eight patients had grade 1 and seven patients had grade 2 anemia. Only two patients had grade 3 anemia, including the patient who died of pancytopenia. Five patients received RBC transfusions on the trial, with four patients receiving only one transfusion and one patient requiring packed cells on three occasions. Other hematologic toxicity was not seen in the patients either on cycle one or cumulatively, except in one patient at the dose level of 2.8 mg/m2 who had grade 4 neutropenia and grade 4 thrombocytopenia. Some mild transient increases in liver enzyme levels (specifically, AST and gamma-glutamyltransferase), were seen at the 2. 4- and 2.8-mg/m2 levels. Two patients had persistent elevations of these enzymes because of disease progression.
Two deaths occurred during the study. One woman with metastatic breast cancer was treated at the highest dose level of 2.8 mg/m2. At the time of her enrollment, her hemoglobin count was 103 g/L, her WBC count was 4.0/dL, her granulocyte count was 3.7/dL, and her platelet count was 100,000/dL. She was admitted 4 days after the infusion with fever, marked clinical deterioration, grade 3 neutropenia, and grade 4 thrombocytopenia. Her bone marrow showed a hypoplastic state, and despite intensive supportive measures, she died 15 days after the infusion. A direct correlation with either drug toxicity or rapidly progressive disease could not be elucidated. The other death occurred in a patient with renal cell carcinoma who died of progressive disease 1 week after receiving the first dose of ONCO-TCS at the 1.0-mg/m2 dose level.
Therapeutic Response
One partial response was defined by computed tomography (CT) scanning of a man with pancreatic carcinoma who received five cycles at the 2.0-mg/m2 dose. This partial response was accompanied by a reduction in pain and analgesic requirements and was manifested by a more than 50% decrease in the size of the mass from the first scan that was maintained for 87 days. Two other patients in the study experienced decreases in the tumor dimensions for less than the 4 weeks required to meet the definition of a partial response. One patient had a lymphoma and was initially treated at the 2.8-mg/m2 dose and subsequently at 2.4 mg/m2 for two doses. The other patient had adrenocortical carcinoma and was treated at the 2.4-mg/m2 dose for two cycles and then at 2.0 mg/m2 for two doses. Three patients at study completion had stable disease, including one patient with each of the following malignancies: a lymphoma, a renal cell carcinoma, and a non–small-cell lung cancer.
Pharmacokinetic Study
For 24 of the 25 patients enrolled in the study, plasma samples were collected for determination of the pharmacokinetic behavior of this agent. Plasma vincristine concentrations determined in this study reflected total drug (both liposome entrapped and nonencapsulated).21 Quantification of nonliposomal vincristine could not be performed because of the extremely low concentrations of free drug that arise from the slow leakage of vincristine from the liposomes in the circulation. These levels were significantly below the limit of quantitation for the HPLC method used to detect vincristine.
Initial results from patients with complete pharmacokinetic data were recently presented.20 These data indicate that the plasma elimination of vincristine after injection of ONCO-TCS is best described by a two-compartment model, suggesting that the liposomes circumvent the initial rapid distribution phase experienced with conventional (nonliposomal) vincristine. Dramatically elevated plasma vincristine concentrations after ONCO-TCS injection, compared with those previously observed for free drug administration, were revealed in both Cmax and AUC pharmacokinetic parameters at doses of 1.0 mg/m2 or higher, as shown in Fig 1A and 1B. Although significant variability in vincristine concentration versus time curves were observed at each dose level, there was an increase in Cmax with increasing dose (Fig 1A) that was statistically significant. In comparison, although a trend for increasing AUC with increasing drug dose for levels of 2.0 mg/m2 and higher was seen (Fig 1B), the variability within dose levels resulted in a lack of statistical significance in this trend.
Dose dependence of plasma maximum vincristine concentrations (A) and AUC vincristine exposure (B) in patients receiving ONCO-TCS. Parametric Cmax and AUC values were determined by methods described in the Patients and Methods section, under Pharmakinetic Sampling. Error bars represent ± 1 standard deviation. {/GRAPH;jcoj01416001bx;80256n;;85536n;0n}
In a study of free vincristine, the whole blood concentration measured at 210 minutes was approximately 15 nmol/L (12 ng/mL).22 In the present study, levels were assessed at approximately the same time or later in patients receiving doses of 2.0, 2.4, or 2.8 mg/m2. The mean vincristine concentrations were 2.5- to 154-fold higher (mean, 547 ng/mL; range, 28.6 to 1,850 ng/mL).
DISCUSSION
Although the greatest benefit provided by liposome encapsulation of vincristine may ultimately be related to increased antitumor potency, the primary goal of this phase I study of liposomal vincristine was to define the pattern and dose response of toxicity and identify a dose that could be recommended for subsequent phase II testing where its therapeutic activity can be evaluated. The results from this clinical trial indicate that increased doses of vincristine can be administered when encapsulated in appropriate liposome formulations compared with conventional vincristine. Specifically, this single-institution study suggests that 2.4 mg/m2 is the MTD and that 2.0 mg/m2 over a 1-hour infusion every 3 weeks is a well-tolerated dose that can be used in subsequent studies.
This phase I study of ONCO-TCS was initiated at a dose of 0.5 mg/m2, which is approximately three-fold lower than the recommended dose for free (nonliposomal) vincristine. This conservative starting dose was selected based on the fact that although ONCO-TCS seemed to exhibit comparable or reduced toxicity compared with free drug in preclinical models, the dose-limiting toxicity in animals was gastrointestinal in nature rather than neurotoxicity, which is dose-limiting in humans. Vinca alkaloid–induced neurotoxicity is difficult to predict in preclinical studies, particularly neuropathy and neuropathic pain. Several attempts were made to compare the neurotoxicity of ONCO-TCS with that of free vincristine in rodent, dog, and avian models, using various dosing schedules. However, in each case, dose-limiting gastrointestinal toxicity was experienced before the onset of any evaluable neurotoxicity (L. Mayer, unpublished observations). In view of the significant increases in neurotoxicity that have been associated with attempts to increase vincristine dose-intensity through removing the 2-mg total dose cap or administering the drug as a multiple-day continuous infusion, a decision was made to initiate the trial at a dose well below that currently used for free vincristine. In addition, the dosing schedule used was once every 3 weeks to avoid any potential cumulative neurotoxicity. The rationale for the dosing regimen selected was based on our prediction that liposome encapsulation would significantly increase the circulation longevity of vincristine. Although these considerations were appropriate in the design of this first phase I trial of liposomal vincristine, the results indicate that the liposome formulation used here does not exacerbate vincristine toxicity, nor does it seem to cause significant cumulative toxicity on an every-3-weeks dosing schedule.
The dose-limiting toxicity was pain and obstipation, the latter likely related to an autonomic neuropathy secondary to the vincristine.12,23 Mild to moderate fevers, rigors, diaphoresis, and headaches during the infusion were seen acutely. Nausea, fatigue, myalgias, and peripheral neuropathy were reported at follow-up visits. There was a clear correlation of increasing toxicity, in both frequency and degree, with increasing dose. The spectrum of toxicities observed with ONCO-TCS correlated to some degree with that of vincristine in that neurotoxicity was one of the dose-limiting toxicities whereas myelosuppression did not seem to be a dose-limiting side effect.23 The pattern of neuropathy was predominantly autonomic and was characterized by constipation and obstipation. This pattern is also observed with other vinca alkaloids. The peripheral neuropathy typically seen with chronic free vincristine dosing was not as pronounced with ONCO-TCS. This less pronounced peripheral neuropathy could be attributed to the small number of drug doses that the majority of patients received in this study; alternatively, it may be related to a somewhat different pattern of toxicity with the liposomal formulation. Similar neurologic toxicities were observed in one study in which free vincristine doses above the 2-mg cap were tested in patients,12 suggesting that the shift to abdominal symptoms may be related to the elevated doses delivered with the liposomal formulation. In regard to the other toxicities observed in this study, the manifestations of fever, rigors, and myalgia seemed related to the liposome formulation of vincristine, as these side effects are not usually observed with administration of vinca alkaloids.
Significant variability was observed in the pharmacokinetic behavior of ONCO-TCS, as evidenced by the wide range of AUC values for most patient cohorts. The degree of variability in this phase I trial is comparable to that reported in the literature for the free drug and could not be attributed to any particular patient characteristic, co-administered medication, or patient dose level.21 Although the relative variability in vincristine elimination was similar for free and liposomal drug, the pharmacokinetic data here demonstrate that the doses of vincristine in liposomal form gave significantly increased AUC and Cmax values. Total systemic exposure (as reflected by AUC values) was as much as 150-fold higher than that previously reported for free vincristine. As described previously, the increased systemic exposure of vincristine when administered in liposomal form seems to be caused by the ability of the liposomes to avoid the initial rapid elimination phase associated with conventional vincristine pharmacokinetics.21
Although there were indications of a trend for increased toxicity with increased Cmax and AUC, no clear correlation between the pharmacokinetic parameters and grade of toxicity for individual patients could be established. This lack of a clear correlation may be due in part to the fact that vincristine concentrations measured here reflect both free and encapsulated vincristine and the liposome entrapped drug would not be expected to cause direct toxicity. After injection of liposomal vincristine, the levels of unencapsulated drug in plasma were far below the limits of quantitation for our HPLC assay. Therefore, we were unable to determine plasma concentrations of free vincristine. A much more sensitive assay, such as one based on radioimmunologic detection, may be required to elucidate the individual contributions of free and encapsulated vincristine on toxicity.
This phase I study of ONCO-TCS highlights some of the difficulties inherent in toxicity studies. The initial patients enrolled at the 2.4-mg/m2 dose level had no significant toxicity, so the dose was increased and patients were enrolled at the 2.8-mg/m2 dose level. The toxicity response observed for that dose fit the definition for the MTD. Subsequently, when three additional patients were enrolled at the 2.4-mg/m2 dose, dose-limiting toxicity was also noted at this level. Therefore, the recommended dose level for phase II studies that emerged from our trial was 2.0 mg/m2, which is, by definition, one dose lower than the level at which dose-limiting toxicity was observed.
The variability in tolerance noted above was not associated with any differences in the formulation characteristics over time and highlights the relative nature of the MTD in phase I trials. When dose-limiting toxicity is defined by biochemical or hematologic changes, there may be more uniformity, but even then it may depend upon exposure to previous chemotherapy, bone marrow tolerance, or heterogeneity of drug metabolism. In this case, in which the MTD was defined as evidence of neurotoxicity, the tolerance was quite variable with the 2.4-mg/m2 dose, causing few problems in some patients but significant side effects in others. These difficulties in defining the MTD are of interest in designing phase I studies. In this study, we used a modified Fibonacci scheme and enrolled three patients on each dose level. In many of our current phase I studies, we are using a continuous reassessment method to decrease the number of patients enrolled at potentially ineffective doses.24 One concern with this approach is the risk that toxicity will be missed if fewer patients are enrolled at each dose level. In reviewing this trial, it seems unlikely that we would have missed toxicities if we had used a continuous reassessment method, arrived at a different MTD, or had less difficulty defining the MTD. We would have still enrolled six or more patients at the higher doses but would have decreased the number of patients at the initial dose levels.
At the highest and recommended phase II dose levels, the total dose of vincristine administered to patients was significantly higher than the standard vincristine dose, especially compared with protocols in which the total dose per patient is capped at 2.0 mg. One of the patients treated with the 2.8-mg/m2 dose received a total dose of vincristine of 6.2 mg on his first treatment cycle. In fact, at all but the lowest dose levels (0.5 and 1.0 mg/m2), patients received total doses of more than 2.0 mg. Preclinical studies indicated that dose escalations with ONCO-TCS were associated with a greater degree of antitumor efficacy.13-15 Whether the level of dose escalation achieved in these study patients will be adequate to observe enhanced efficacy in human tumors will not be known until phase II studies in potentially sensitive tumors are completed. In this context, it is of interest to note that therapeutic activity of ONCO-TCS was observed in this trial, with one confirmed partial response and other indications of antitumor activity in a variety of tumors. Response is not the major end point of a phase I study, but the suggestion of therapeutic efficacy is promising. When one reviews the historical data with vincristine, responses have been reported in a wide number of tumor types, including colon cancer, breast cancer, sarcoma, lung cancer, lymphomas, and leukemias.25 We observed a partial response in pancreatic cancer, a tumor that conventionally responds poorly to cytotoxic agents and where responses are often not observed on CT scanning with agents considered active. The CT scan-confirmed response in this study suggests that further investigation of other patients with pancreatic cancer is warranted.
In summary, this study confirms that ONCO-TCS can be safely administered as a 1-hour infusion every 3 weeks at doses significantly higher than that typically used for free vincristine. The dose-limiting toxicity was neurotoxicity with pain, obstipation, and peripheral neuropathy, while therapeutic activity was seen with one partial response. Further studies of ONCO-TCS have been initiated to evaluate the potential utility of this formulation in specific clinical settings.
Acknowledgments
Supported in part by Inex Pharmaceuticals, Vancouver, BC, Canada.
- Received July 7, 1998.
- Accepted October 20, 1998.
